Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

ASML became the chokepoint for cutting-edge chips

worksinprogress.co

Post reply on HN
Print view View original

221–230 of 251 posts

Re: ASML became the chokepoint for cutting-edge chips

#221

Post by shmeeed »

robocat wrote: post #218
ghc wrote: post #113

Earlier quoted context omitted.

Let's not forget the really big gamble (inventing EUV) was made in the 1990s by US national labs: Sandia, Lawrence Livermore, and Lawrence Berkeley.

Every country tries to claim critical dependency for innovations. I guess you parochially mention that one because you remember that narrative (maybe because it makes you feel better). I maliciously searched for examples of the French thinking they invented the transistor. Turns out the French do have a claim of inventing it simultaneously (at the same time as Bell) and the French even commercialised their version si…

>Apparently it was two German inventors working for Westinghouse in France on a project for French telecoms as I recall

So the Germans invented it! JK

I like how you put this. It always seems weird to me how some people get hung up on these claims when it's so obvious that history is full of basically simultaneous inventions.

Re: ASML became the chokepoint for cutting-edge chips

#222

Post by ghc »

robocat wrote: post #218
ghc wrote: post #113

Earlier quoted context omitted.

Let's not forget the really big gamble (inventing EUV) was made in the 1990s by US national labs: Sandia, Lawrence Livermore, and Lawrence Berkeley.

Every country tries to claim critical dependency for innovations. I guess you parochially mention that one because you remember that narrative (maybe because it makes you feel better). I maliciously searched for examples of the French thinking they invented the transistor. Turns out the French do have a claim of inventing it simultaneously (at the same time as Bell) and the French even commercialised their version si…

Except ASML licensed the technology from the US Government, after the government labs built the first EUV fab in 2001. Not to take anything away from ASML...all the US companies that also licensed the tech failed to commercialize it, but the US Government blocked Canon and other Japanese companies from acquiring the technology. The entire reason ASML has the technology and nobody in Japan does is geopolitical.

Re: ASML became the chokepoint for cutting-edge chips

#223

Post by vel0city »

whilenot-dev wrote: post #183
andsoitis wrote: post #168

Earlier quoted context omitted.

A machine is a device that uses energy to perform work. Typically by applying or transforming force, motion, or both. The space shuttle can be thought of maybe as a collection of machines working in concert, but thinking of it as ONE machines renders the meaning of machine less useful.

In my understanding a device has its origin in giving advice and does something specific for you (a pen is a writing device, a mixer is a cooking device, a phone is a communication device, a bus is a transportation device etc.). A machine on the other hand has its roots in its mechanisms. It physically transforms something by applying mechanical power, and that's not necessarily done for you (e.g. printing device VS…

Device comes from Latin dividere, meaning "something which is divided". Later with old French devis (disposition, desire, purpose, or decorative emblem) and deviser (arrange, plan). A device is something planned, designed, potentially intricate. A device doesn't always need to be mechanical/physical as there are "literal devices" and one can be "left to their devices". I'd say "device" is more like "a planned thing" if giving a basic definition.

A machine is almost always a device, but a device isn't always a machine. A fancy earring can be a device, but it is clearly not a machine.

Re: ASML became the chokepoint for cutting-edge chips

#224

Post by pjc50 »

quanto wrote: post #133

I have an interesting discussion with a senior colleague: why ASML? why are they by far the best? Their competitors are a few generations behind. The colleague claimed that there is no special magic. It's not that ASML is using some otherwise unknown laws of physics nor is any single step or component particularly special or novel. It's just that they meticulously optimized each step, and the sum of such steps is the…

> It's not that ASML is using some otherwise unknown laws of physics nor is any single step or component particularly special or novel. It's just that they meticulously optimized each step, and the sum of such steps is the winning solution.

Previously in the context of Apple I likened this to becoming a chess grandmaster: all you have to do is make the optimal decision every time you make a move, over and over again, for years

People don't like hearing that there isn't One Weird Trick which you can just copy, but it's the reality of these situations. To the extent that they can be analyzed, the best people to send are often anthropologists to look at the decision making culture. Culture is even harder to copy; this was a factor in the difficulties of TSMC Arizona starting up, despite it being literally the same company it's not the same people.

Re: ASML became the chokepoint for cutting-edge chips

#225

Post by pjc50 »

jstummbillig wrote: post #25

> By betting on extreme ultraviolet lithography long before it worked, ASML became the chokepoint for cutting-edge chips. Makes one wonder: Would we be much better off of worse off if we reshaped society to do more of things, where a new technology is unlikely to work but highly beneficial in the limits? Would we sooner have 10 additional ASMLs or waste a lot of resources?

> Would we sooner have 10 additional ASMLs or waste a lot of resources?

It seems obvious that the answer would be both? All of these things are "bets", at almost every level.

What happens is someone comes along and notices the 9/10 failed attempts and decides that the right thing to do is only to make bets which are guaranteed to win. So they get smaller and smaller.

Re: ASML became the chokepoint for cutting-edge chips

#226

Post by pjc50 »

hackrmn wrote: post #28

> These machines are roughly the size of double-decker buses. To ship one requires 40 freight containers, three cargo planes, and 20 trucks. They are the world’s most complex objects. Each contains over one hundred thousand components, all of which have to be perfectly calibrated for the machine to produce light consistently at the right wavelength. As a software engineer by trade, the above parable communicates to m…

> reducing the complexity and commoditising the whole thing is what's really going to be impressive I think

What do you think "cutting edge" is, or Moore's law has been?

At one point you could have written a similar article about, say, 165nm, which is now going to the scrapyard. In the past these things have always gradually got more available and easier, with higher yields - but a new, better one appears.

But at some point we're going to reach an equilibrium with physics itself. Where, even with all the complexity we can muster, it's not possible to make it easier or get smaller.

Re: ASML became the chokepoint for cutting-edge chips

#227

Post by sehansen »

locallost wrote: post #182
CGMthrowaway wrote: post #164

Earlier quoted context omitted.

Carl Zeiss Meditec /= Carl Zeiss. $AFX is the publicly traded non-semis division. Carl Zeiss AG is the parent company which is private

Correct. Funny enough though, their corporate structure and the name AG means they do have stocks, but they are not traded and 100% privately owned. For some reason I see this often with German companies, e.g. the German railway. Not sure why that is, although for the railway plausible since they are owned by the state that might eventually want to sell parts of it.

An AG corporation has stocks in order to track who owns how much and also attach different economic and voting rights to different classes of stock. The other way to incorporate a limited-liability company in Germany is the GmbH, which tracks ownership directly in the articles of incorporation, but are in other ways subject to way lower management, disclosure and accounting requirements. So the AG is mostly useful if you want it to be easy to change your ownership structure, if you for instance raise capital from new investors, issue employee shares, change cross-ownership within a conglomerate or go public some day.

Why Carl Zeiss is an AG I don't know. The West German Carl Zeiss was re-formed as a GmbH in 1946, but had changed to an AG by 1973. The East German Carl Zeiss was turned into a GmbH during reunification and then split in two. One part merged into the West German Carl Zeiss AG and the other is now called Jenoptik. Jenoptik was converted into an AG in 1996 and went public in 1998. AFAICT Carl Zeiss has been privately owned by the Carl Zeiss-Stiftung since 1889, except of course for the temporary East German part.

Re: ASML became the chokepoint for cutting-edge chips

#229

Post by alnwlsn »

robotresearcher wrote: post #118
kens wrote: post #82

Is the ASML machine actually the world's most complex machine under some metric, or is this a claim that someone made up? I.e., did someone actually compare the ASML machine to the Space Shuttle, LHC, Internet, and so forth and show that it is more complex under some definition? (I've done various historical questions, so I'm sensitive to how statements are sourced.) An orthogonal question is what makes sense as a me…

2.5 million moving parts sounds unlikely. Parts, sure. Moving parts, I’m skeptical.

I guess that depends on what is considered moving. A latch is, but is a vibration dampener? On launch, I bet a lot of parts are moving that you probably would not otherwise want to be.

Re: ASML became the chokepoint for cutting-edge chips

#230

Post by cma »

joe_the_user wrote: post #161
quanto wrote: post #133

I have an interesting discussion with a senior colleague: why ASML? why are they by far the best? Their competitors are a few generations behind. The colleague claimed that there is no special magic. It's not that ASML is using some otherwise unknown laws of physics nor is any single step or component particularly special or novel. It's just that they meticulously optimized each step, and the sum of such steps is the…

Technology can explain why ASML or other companies are dominant at a given moment but it can't explain how it became dominant. Essentially, technological superiority involves research and that requires money. IE, it is a product of investment in the various technologies. The history of ASML involves a "failed" company that other multinationals felt they had to keep alive to allow the technologies to continue. And tha…

> it being in Europe is perhaps a combination of European tradition and Europe wanting some part of the global chip production system

The US doesn't randomly hold export veto power on ASML through unilateral threats. EUV had a lot of tech transfer from the US of the initial research as the sibling comment lays out, and the agreements for those transfers allowed restrictions.

Post reply on HN